Site Home   Archive Home   FAQ Home   How to search the Archive   How to Navigate the Archive   
Compare FPGA features and resources   

Threads starting:
1994JulAugSepOctNovDec1994
1995JanFebMarAprMayJunJulAugSepOctNovDec1995
1996JanFebMarAprMayJunJulAugSepOctNovDec1996
1997JanFebMarAprMayJunJulAugSepOctNovDec1997
1998JanFebMarAprMayJunJulAugSepOctNovDec1998
1999JanFebMarAprMayJunJulAugSepOctNovDec1999
2000JanFebMarAprMayJunJulAugSepOctNovDec2000
2001JanFebMarAprMayJunJulAugSepOctNovDec2001
2002JanFebMarAprMayJunJulAugSepOctNovDec2002
2003JanFebMarAprMayJunJulAugSepOctNovDec2003
2004JanFebMarAprMayJunJulAugSepOctNovDec2004
2005JanFebMarAprMayJunJulAugSepOctNovDec2005
2006JanFebMarAprMayJunJulAugSepOctNovDec2006
2007JanFebMarAprMayJunJulAugSepOctNovDec2007
2008JanFebMarAprMayJunJulAugSepOctNovDec2008
2009JanFebMarAprMayJunJulAugSepOctNovDec2009
2010JanFebMarAprMayJunJulAugSepOctNovDec2010
2011JanFebMarAprMayJunJulAugSepOctNovDec2011
2012JanFebMarAprMayJunJulAugSepOctNovDec2012
2013JanFebMarAprMayJunJulAugSepOctNovDec2013
2014JanFebMarAprMayJunJulAugSepOctNovDec2014
2015JanFebMarAprMayJunJulAugSepOctNovDec2015
2016JanFebMarAprMayJunJulAugSepOctNovDec2016
2017JanFebMarAprMayJunJulAugSepOctNovDec2017
2018JanFebMarAprMayJunJulAugSepOctNovDec2018
2019JanFebMarAprMayJunJulAugSepOctNovDec2019
2020JanFebMarAprMay2020

Authors:A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Custom Search

Messages from 22475

Article: 22475
Subject: Re: Error with Quartus for Altera APEX20K device: clock skew is greater then data delay
From: "Xanatos" <deletemeaoe_londonfog@hotmail.com>
Date: Wed, 10 May 2000 02:10:01 GMT
Links: << >>  << T >>  << A >>
> I am a novice user of the Altera software and I am stuck with a design
that
> I run through Quartus 2000.03 and I get an error during compilation: clock
> skew is greater then data delay, ciruit will not function.  If this is

Is this on extra effort? Did you try to assign a clique to a block? Are your
setup/hold times set? Check to ensure that the settings in the Assignment
Dis-Organizer are configured properly for your design.
Also, did you specify a global clock, or did you "specify" the clock?

Check that, and if it still gives you grief, let me know.

Cheers,
Xanatos


Article: 22476
Subject: Xilinx Student Edition 1.5 License.dat
From: "R. T. Finch" <robfinch@cyg.net>
Date: Tue, 9 May 2000 22:26:43 -0400
Links: << >>  << T >>  << A >>
I have the following problem trying to run synthesis:

FPGA Express Macro Compiler
Version 3.1.1.0w

Initialize DPM...

Checking license...

Checking license for Synopsys failed.


I received and set up a license.dat file according to the supplied
instructions, but apparantly it doesn't work. I am using the license.dat
file (verbatium) as e-mailed to me by the online registration. For some
reason they emailed me two attachments, one with linefeeds and one without.
neither works.

Since Xilinx does not offer live support I am unable to contact them, and I
am unable to make use of their software. This is very aggravating. None of
the problems in their FAQ or database provide a resolution to this problem.

Please help.

Thanx
Rob



Article: 22477
Subject: Re: Looking for Altera programmer in France
From: Ashok Mahadevan <ashokm1@earthlink.net>
Date: Wed, 10 May 2000 02:39:04 GMT
Links: << >>  << T >>  << A >>
> Hi
> I have an unusual request: we have a DataIO programmer that can't
> program our Altera PROMs. We don't want to invest in an other programmer
> so we are looking for someone/a company/anything else who owns an Altera
> programmer near us (Courbevoie, France) who could program the PROMs for
> us.
> Thanks in advance

Did you try the Altera distributors (Arrow?) themselves? They can usually
pre-program the PROMs for you.....

Ashok

Article: 22478
Subject: Re: ANNOUNCE: Embedded Systems Glossary and Bibliography
From: OneStone <onestone@chariot.net.au>
Date: Wed, 10 May 2000 12:50:26 +0930
Links: << >>  << T >>  << A >>
Rennie Allen wrote:
> 
> OneStone wrote:
> >
> > I probably just move in much older circles than you,
> 
> Hmmm, I'm feeling pretty old right now, thanks for making me feel
> younger ;-)
> 
> > before the computer
> > world went insane with jargon.
> 
> What year, exactly, was that ?  Was that the year Grace Hopper coined
> the term "bug" ?  Seems to me, that any substantial technology will
> inevitably require its own terminology.  

And isn't it interesting trying to keep up with it all, hence the
glossary is a great idea.

> I have no problem with this, > I
> simply want whatever terminology that is developed to be clear, and
> concise (btw: I don't consider "bug" to be in the category of clear and
> concise, although it is quaint).

Bug is so well known it now transcends culture, as in different
technical backgrounds. While mutex doesn't as can be seen from other
replies. To each his own, my comment is simply that to include omne
term, and omit another leaves the glossary incomplete.

> 
> > Mutual exclusion mechanism smacks of the
> > same sort of mentality that turned housewifes into domestic engineers,
> > and directory enquiries operators into information consultants. Since
> 
> I disagree.  MUTual EXclusion is the function of the device.
> Personally,
> I find functional naming clearer than say, the term "bit flag" (which is
> incredibly overloaded, and gives no clue as to the intended function of

Flags are not necessarily bits, although the term is most often used in
this way. I have used it for so long that I really have never felt the
requirement for anything else.

> the
> device - and in no way conveys the requirement for atomic state change).
> 
> Contrast the rationale behind this, to that behind the term "domestic
> engineer".  The term "domestic engineer" is more ambiguous than
> housewife.
> I contend that mutex is *less* ambiguous than "bit flag", since a mutex
> can only be one thing, while a "bit flag" could be many things (much as
> housewife can mean only one thing, while "domestic engineer" could mean
> many things).

I agree, sorry, poor comparison, although mutex only conveys meaning if
you already fully understand the context, it might not be as meaningful
when trying to explain soemthing to a mildly techniocal client. 

> 
> > pthreads are a POSIX multithreading standard I have never had cause to
> > read about them extensively or use them in any embedded systems. And a
> > review in Amazon showed that only the Butenhof book seemed to be really
> > popular, the numbers sold on this subject hardly makes the book
> > 'popular'.
> 
> Hmmm, Amazon.com, as dearly as they may want us to believe so, is not
> the
> only source of technical reference material on the planet, hence your
> data
> (with regards to popularity) may be incomplete.

I did check through 2 mmajor national book sellers as well, and one of
the university book stores. The only one showing any real activity, as
might be expected, was the uni store, even then, from an overall
perspective of book sales related to programming it could not be termed
exactly popular.

> 
> Certainly, I didn't mean to imply that these books were popular relative
> to,
> say Stephen King novels :-)
> 
> > Perhaps reasonably popular amongst POSIX programmers, but my
> 
> A reasonable guess would be that every real-time Posix programmer, has a
> copy
> of at least one pthreads book.
> 
> > HC12, HC16, and even my MCORE systems aren't running POSIX, definitely
> > none of my 8 bitters do.

I guess my response was too related to comp.arch.embedded, where I read
it. I'm not aware of too many embedded POSIX systems.

> 
> Understood.  My original point was simply that I don't believe the
> Embedded
> Systems Glossary was being obscure, simply by including the term mutex;
> in
> fact, it appears to be a necessary inclusion (since the purpose of a
> glossary
> is to explain terms).  Your point was that the term "bit flag" should be
> included, and while this would be nice, I can understand why someone
> attempting
> to produce a concise glossary would avoid this term (due to the fact it
> is
> extensively overloaded).

But then the glossary is incomplete without it. The author asked for
comment, I sent a few comments to him, hopefully in a constructive
manner, not suggesting that any term be excluded, in fact the very
existence of the glossary has forced me to go and have a look at a
couple of items which were known to me, but not in everyday usage. I
simply suggest that a glossary of embedded systems programming, which I
believe this is, should include as many terms as are in common use. I
personally feel that flag (and semaphore) are more common amongst
embedded systems programmers, I could of course be wrong, I often am.

Al
Article: 22479
Subject: Re: ANNOUNCE: Embedded Systems Glossary and Bibliography
From: OneStone <onestone@chariot.net.au>
Date: Wed, 10 May 2000 12:57:46 +0930
Links: << >>  << T >>  << A >>
PeterS wrote:
> 
> >> I don't like the use of "flag" in such specific terms - to me, a flag is
> >> simply an
> >> indication of state, with no protection mechanism for setting or
> resetting.
> >
> >The original system of semaphores, for signalling, used flags. A
> >semaphore is simply a signal, as is a flag, flags are not necessarily
> >binary, originally semaphore, the method of singalling, each with
> >different context sensitive meaning.
>

Them ancient fingers don't work so well at 3am any more. To correct my
cock up:-

the method of signalling used several flags of different designs, each
with different context sensitive meaning.
 
> Hmmm, interesting sentence. Wonder what it means ;-)
> 
> Chains of simple signal flags were also used to "post" entire messages from
> the rigging
> of sailing ships (like Nelson's "England expects this day..." signal at
> Trafalgar),
> but there was nothing, I guess to stop anyone else (ie no interlock) putting
> up a
> message from their own rigging, which is why I dislike the interchangeable
> use of
> "flag" and "semaphore".

The interlock, in Napoleonic times was the yard arm, as dispensed by
british maritime law, disobey the signal and suffer for it. I guess that
makes Fletcher Christian a 'bug'.

Al

> 
> >I believe it originated in the 18th century.
> >
> >Semaphore  noun [U]
> >             a system of communication using two mechanical arms or
> >             hand-held flags which are moved into different positions to
> >             represent different letters, numbers or symbols
> >             Semaphore was widely used at sea, before the advent of
> >             electricity.
> >
> >             (figurative) When I lived opposite her we would send
> >             semaphore signals (=messages without speaking) to each
> >             other from our bedroom windows.
> >
> >Al
> 
> I live near to a restored semaphore tower dating from the time of the
> Napoleonic wars
> which signalled from the Admiralty in central London down a chain of towers
> to
> Portsmouth and Plymouth.( Just off the A3 outside the M25.)
> It has working arms, and an interesting display model of an earlier
> system which used six flat, square shutters which could be individually
> displayed
> upright, face-on, or rotated horizontal. Somewhat OT, I realise, but a
> fascinating
> monument. Oh yes, a short message took less than 15 minutes from London to
> Portsmouth, weather permitting (50 odd miles, 80+ km).

Amazing how these ancient systems are faster thean the net can be some
days ;@}
Article: 22480
Subject: Re: ANNOUNCE: Embedded Systems Glossary and Bibliography
From: OneStone <onestone@chariot.net.au>
Date: Wed, 10 May 2000 13:02:07 +0930
Links: << >>  << T >>  << A >>
Rennie Allen wrote:
> 
> PeterS wrote:
> >
>  >
> > >The original system of semaphores, for signalling, used flags. A
> > >semaphore is simply a signal, as is a flag, flags are not necessarily
> > >binary, originally semaphore, the method of singalling, each with
> > >different context sensitive meaning.
> >
> > Hmmm, interesting sentence. Wonder what it means ;-)

Hung the flags upside down. ;@}

> >
> > Chains of simple signal flags were also used to "post" entire messages from
> > the rigging
> > of sailing ships (like Nelson's "England expects this day..." signal at
> > Trafalgar),
> > but there was nothing, I guess to stop anyone else (ie no interlock) putting
> > up a
> > message from their own rigging, which is why I dislike the interchangeable
> > use of
> > "flag" and "semaphore".
> 
> I agree absolutely.  IMO, clarity is not served by overloading existing
> terms
> with new meanings.  Sometimes, it is simply necessary to create a new
> word to
> describe a specific concept.
> 
> Rennie

But this is exactly one of the arguments used in favour of C++ and other
object oriented languages, here overloading is portrayed as a 'good'
thing. I don't agree, the original post was not about likes and
dislikes, simply completeness of a glossary.
Article: 22481
Subject: Re: ANNOUNCE: Embedded Systems Glossary and Bibliography
From: OneStone <onestone@chariot.net.au>
Date: Wed, 10 May 2000 13:03:17 +0930
Links: << >>  << T >>  << A >>
Rennie Allen wrote:
> 
> David Brown wrote:
> >
> > Johan Kwisthout wrote in message <3917fdf7.20732581@obsserver>...
> > >On Mon, 08 May 2000 22:30:04 -0400, Jerry Avins <jya@ieee.org> wrote:
> > >
> > >>Result of check: I asked a CS friend (he has a masters degree and is
> > >>active in the profession) if he knew or could guess what a mutex is. He
> > >>had a vague memory that it is a species of mosquito.
> > >>
> > >>Jerry
> > >
> > >Funny, I study CS in the Netherlands and the theory of operation
> > >systems (like synchronisation algorithms, multitasking theory) was
> > >part of the first half of the bachalors program, dealing with
> > >semaphores, mutexes, critical regions, producers/consumers, deadlocks
> > >etcetera. Guess it depends on the university where you're studying...
> > >
> >
> > It certainly does.  Some universities and courses will cover things like
> > Windows programming, which does not really have anything to do with CS, but
> > which looks good to PHBs for later employment.
> 
> Are you implying, that mutex is MS-speak ?  It isn't (although the
> ingrates
> do use the term .... there should be a law .... ;-)
> 
> Rennie


There is, and finally they've (michael soft) run foul of it, or so it
seems. ;@}
Article: 22482
Subject: Re: ANNOUNCE: Embedded Systems Glossary and Bibliography
From: Rennie Allen <rennieallen@earthlink.net>
Date: Wed, 10 May 2000 05:00:51 GMT
Links: << >>  << T >>  << A >>
OneStone wrote:
> 
> Rennie Allen wrote:

> 
> But this is exactly one of the arguments used in favour of C++ and other
> object oriented languages, here overloading is portrayed as a 'good'
> thing.

Hmmm, I've seen a little too much of this "good" thing in C++ code. 
Inappropriate semantic overloading, is one of the most abused facilities
of C++ IMO.

> I don't agree, the original post was not about likes and
> dislikes, simply completeness of a glossary.

Definately off topic.

Rennie
Article: 22483
Subject: Re: HELP - what to choose?
From: Rickman <spamgoeshere4@yahoo.com>
Date: Wed, 10 May 2000 01:14:43 -0400
Links: << >>  << T >>  << A >>
Ray Andraka wrote:
> 
> If you look at the FPGA architectural features as they apply to a computer
> design, I think you'd wind up settling on the Xilinx architectures.  First,
> you'll probably want a fast carry chain so that you can do fast arithmetic
> with minimal logic.  That rules out pretty much everyone except Xilinx,
> Altera and Lucent.  I haven't seen any commercial boards with Lucent, so bye
> bye Lucent.  Xilinx has a capability of using a LUT (a four input look-up
> table which is the basis of the FPGA architecture) as a small RAM or shift
> register.  This makes for very compact register files, pipeline delays,
> reordering queues etc.  In Altera, each bit of storage chews up another
> logic element unless you use the buld memories.  Also, Altera's carry logic
> structure is not as powerful as the Xilinx structure, which means you will
> probably need two levels of logic for real-world arithmetic vs 1 level of
> logic in Xilinx.
> 
> If you want it small, you could always go to a bit serial design.  A pretty
> decent  bit serial scientific calculator will fit in an XCS-05

Hey Ray! What about the PC104C31 which uses not only one, but four
Lucent FPGA chips? I realize that our price does not fit the $300 limit
imposed in this case, but don't say there aren't any commercial boards
around that use the Lucent chip! :) 

Of course our board is not really intended to be an FPGA evaluation
board, but rather a DSP board with FPGA capability. But it is so loaded
with FPGA capability that anyone needing a lot of FPGA in multiple chips
would do well to look at it and just consider the DSP to be a fancy boot
loader. In fact the OR3T30 or OR3T55 that can be used on our board would
be a very good choice for an attached coprocessor for the TMS320C31 that
is on the board. 

I believe that APS also has a Lucent board much like their APS-X84 board
although they may not be marketing it. They sent me a board at one point
as a beta test which I never got around to using. 

 
> Andy Holt wrote:
> 
> > This is the sort of thing I would expect to be an FAQ, but there doesn't
> > seem to be one for this group.
> >
> > I have been thinking about "playing" with an FPGA both from the view of
> > learning about an interesting-looking technology and with the hope of
> > constructing an emulation of a '60s mainframe (more about this later).
> >
> > I am looking for advice on low-cost ways of doing this (Let's say price
> > ceiling of about £200 [$300]). It seems that I am going to need two main
> > things:
> > * A package of software.
> > * A prototyping/evaluation board.
> >
> > Taking the second of these first there seem to be few choices available
> > (without paying lots of $$$) -

Our board is a lot of money by comparison. The single piece price I have
been quoting is between $2000 and $3000 depending on features and speed.
But this will be changing to something a little more palatable at the
end of the month. And of course the price is much better at quantity. 


...snip...
> > As for software, the Kanda and Atmel packages come with some, for the
> > Xess one I would also have to spend another $100 for the Foundation
> > student edition.

The PC104C31 comes with a boot monitor program to allow you to program
the Flash, boot the FPGAs and perform a lot of testing of the various
parts of the board. Full source is provided along with the designs for
the FPGAs. A low end package for the Lucent FPGAs is available for $150
very much like the introductory package from Xilinx. This package
includes Viewlogic schematic capture and simulation tools. I also
believe it includes VHDL synthesis. 


> > ** so, first question: any known "gotcha's" with the above alternatives?
> >   [ISTR a recent hint that the Atmel software was weak in one respect -
> > it is noticeable that their web site seems to say almost nothing about
> > its functionality - and the low cost version of Foundation doesn't
> > include VHDL?]
> > Are there other reasonable options?
> >
> > The other main question I have concerns estimating how big an FPGA I
> > would need for the mainframe emulation. I assume that the "usable gate"
> > counts for all devices tend to be as much marketing as technical
> > statements. I have detailed (but only "almost complete") descriptions of
> > the logic design for the mainframe that I am interested in (ICT 1905 -
> > aka FP6000) and I can be reasonably confident that it has less than 6000
> > gates including FPU ... probably less than 4000 without. Is this likely
> > to fit in a "10000 gate" FPGA?, a "20000 gate" one, or whatever?

Keep in mind that your gate count also needs to include the IO system
for accessing the memory and any peripherals that will be attached. This
can easily double the size of the design. 

Yes, you are right about the gate counts for the FPGAs. But the largest
part of the gate inflation comes from the inclusion of memory in the
gate counts. They make certain assumptions about the percentage of the
memory used. Your design may use more or less of that memory. This will
change the used gate count dramatically. 


-- 

Rick Collins

rick.collins@XYarius.com

remove the XY to email me.



Arius - A Signal Processing Solutions Company
Specializing in DSP and FPGA design

Arius
4 King Ave
Frederick, MD 21701-3110
301-682-7772 Voice
301-682-7666 FAX

Internet URL http://www.arius.com
Article: 22484
Subject: Re: Xilinx Student Edition 1.5 License.dat
From: Joachim Hoch <joachim.hoch@ipm.fhg.de>
Date: Wed, 10 May 2000 07:33:33 +0200
Links: << >>  << T >>  << A >>
"R. T. Finch" schrieb:

> I received and set up a license.dat file according to the supplied
> instructions, but apparantly it doesn't work. I am using the license.dat
> file (verbatium) as e-mailed to me by the online registration. For some
> reason they emailed me two attachments, one with linefeeds and one without.
> neither works.

One of my colleagues had such a problem when he received a license via email.
The problem were missing \ (backslash) at the end of the lines in the license.
Maybe tjis can solve your problem.

regards

Joachim


--
Joachim Hoch
Fraunhofer Institut Physikalische Messtechnik
Heidenhofstr. 8
D-79110 Freiburg im Breisgau

Tel: +49(0)761/8857-149
Fax: +49(0)761/8857-224


Article: 22485
Subject: Re: pipeline shiftreg in virtex
From: Emil Blaschek <emil.blaschek@siemens.at>
Date: Wed, 10 May 2000 08:30:29 +0200
Links: << >>  << T >>  << A >>
I do not know what problem you have.
if you describe 2 or more consecutive registers without reset 
and only shif funktion, you always get a shift reg, iven if a single FF
and a smallae shift reg would be better.
This works at all versions of FPGA-Compiler II after 1.1.2000

With best regards
E.Blaschek
Article: 22486
Subject: Re: ANNOUNCE: Embedded Systems Glossary and Bibliography
From: OneStone <onestone@chariot.net.au>
Date: Wed, 10 May 2000 16:10:07 +0930
Links: << >>  << T >>  << A >>
Rennie Allen wrote:
> 
> OneStone wrote:
> >
> > Rennie Allen wrote:
> 
> >
> > But this is exactly one of the arguments used in favour of C++ and other
> > object oriented languages, here overloading is portrayed as a 'good'
> > thing.
> 
> Hmmm, I've seen a little too much of this "good" thing in C++ code.
> Inappropriate semantic overloading, is one of the most abused facilities
> of C++ IMO.

Couldn't agree with you more.

> 
> > I don't agree, the original post was not about likes and
> > dislikes, simply completeness of a glossary.
> 
> Definately off topic.
> 
> Rennie

But an embedded systems glossary is on-topic, surely?

Cheers

Al
Article: 22487
Subject: Re: virtex e lvds clock recovery
From: Emil Blaschek <emil.blaschek@siemens.at>
Date: Wed, 10 May 2000 09:14:33 +0200
Links: << >>  << T >>  << A >>
Yes, 
use a "alexander-decoder" for phase recovery 
  either with 3 times oversampling or 
  with 2 consecutive externel delay-lines 
and an external VCO 
   for clock recovery or 
   an alligner

with best regards 
E.Blaschek
Article: 22488
Subject: Re: Xilinx Student Edition 1.5 License.dat
From: Jon Kirwan <jkirwan@easystreet.com>
Date: Wed, 10 May 2000 00:22:26 -0700
Links: << >>  << T >>  << A >>
On Wed, 10 May 2000 07:33:33 +0200, Joachim Hoch
<joachim.hoch@ipm.fhg.de> wrote:

>"R. T. Finch" schrieb:
>
>> I received and set up a license.dat file according to the supplied
>> instructions, but apparantly it doesn't work. I am using the license.dat
>> file (verbatium) as e-mailed to me by the online registration. For some
>> reason they emailed me two attachments, one with linefeeds and one without.
>> neither works.
>
>One of my colleagues had such a problem when he received a license via email.
>The problem were missing \ (backslash) at the end of the lines in the license.
>Maybe tjis can solve your problem.

I received two, as well.  The files use CR LF or else just LF, I
think:  MS-DOS or UNIX defaults.  Probably their software uses either,
so it's more a convenience for those wanting to use TYPE or cat to
display the contents.

Mine appears to work just fine, though I'm no expert on the subject.
One thing to check is the SET variable LM_LICENSE_FILE and make
certain that the license file sits in that same directory.  In my
case, I think that's in fndtn\data.

Jon
Article: 22489
Subject: Re: Xilinx Student Edition 1.5 License.dat
From: Alan Fitch <alan.fitch@doulos.com>
Date: Wed, 10 May 2000 08:54:58 +0100
Links: << >>  << T >>  << A >>
In article <NQ3S4.67488$VR.1345263@news5.giganews.com>, R. T. Finch
<robfinch@cyg.net> writes
>I have the following problem trying to run synthesis:
>
>FPGA Express Macro Compiler
>Version 3.1.1.0w
>
>Initialize DPM...
>
>Checking license...
>
>Checking license for Synopsys failed.
>
>
We use this system on our PCs when we setup courses, and it works. We
normally put the file in 

  c:\flexlm\xilinx.dat

and then add it to the environment variable LM_LICENSE_FILE, e.g.

  c:\flexlm\xilinx.dat;1700@fred

if you have floating licenses on server "fred" as well.

As far as I know, if you put it in the foundation installation
directory, you probably don't need to set LM_LICENSE_FILE. If you put it
anywhere else, you should set LM_LICENSE_FILE

regards

Alan

-- 
Alan Fitch
DOULOS Ltd. 
        Church Hatch, 22 Market Place, Ringwood, BH24 1AW, Hampshire, UK
Tel: +44 (0)1425 471 223                    Email: alan.fitch@doulos.com
Fax: +44 (0)1425 471 573             
**               Visit THE WINNING EDGE  www.doulos.com               **

Article: 22490
Subject: Re: ANNOUNCE: Embedded Systems Glossary and Bibliography
From: "David Brown" <david.nospam@westcontrol.com>
Date: Wed, 10 May 2000 10:04:02 +0200
Links: << >>  << T >>  << A >>

Rennie Allen wrote in message <391835C4.8ACF6356@computermotion.com>...
>OneStone wrote:
>>
>> I probably just move in much older circles than you,
>
>Hmmm, I'm feeling pretty old right now, thanks for making me feel
>younger ;-)
>
>> before the computer
>> world went insane with jargon.
>
>What year, exactly, was that ?  Was that the year Grace Hopper coined
>the term "bug" ?  Seems to me, that any substantial technology will
>inevitably require its own terminology.  I have no problem with this, I
>simply want whatever terminology that is developed to be clear, and
>concise (btw: I don't consider "bug" to be in the category of clear and
>concise, although it is quaint).
>


The term "bug" goes at least as far back as Thomas Eddison - one of his
journals refers to a hardware problem in a design as a "bug".



Article: 22491
Subject: Re: ANNOUNCE: Embedded Systems Glossary and Bibliography
From: "David Brown" <david.nospam@westcontrol.com>
Date: Wed, 10 May 2000 10:10:32 +0200
Links: << >>  << T >>  << A >>

Rennie Allen wrote in message <391837BA.D0B39CFB@computermotion.com>...
>David Brown wrote:
>>
>> Johan Kwisthout wrote in message <3917fdf7.20732581@obsserver>...
>> >On Mon, 08 May 2000 22:30:04 -0400, Jerry Avins <jya@ieee.org> wrote:
>> >
>> >>Result of check: I asked a CS friend (he has a masters degree and is
>> >>active in the profession) if he knew or could guess what a mutex is. He
>> >>had a vague memory that it is a species of mosquito.
>> >>
>> >>Jerry
>> >
>> >Funny, I study CS in the Netherlands and the theory of operation
>> >systems (like synchronisation algorithms, multitasking theory) was
>> >part of the first half of the bachalors program, dealing with
>> >semaphores, mutexes, critical regions, producers/consumers, deadlocks
>> >etcetera. Guess it depends on the university where you're studying...
>> >
>>
>> It certainly does.  Some universities and courses will cover things like
>> Windows programming, which does not really have anything to do with CS,
but
>> which looks good to PHBs for later employment.
>
>Are you implying, that mutex is MS-speak ?  It isn't (although the
>ingrates
>do use the term .... there should be a law .... ;-)
>


Not at all.  MS might use the term (although only a very small proportion of
Windows programmers will ever come across it), but I was implying that you
will meet the term more often in courses with a solid theoritic basis, such
as Johan seems to be studying in the Netherlands, than in simpler courses
covering specific systems (such as Windows) aimed at immediate employment.



Article: 22492
Subject: Re: ANNOUNCE: Embedded Systems Glossary and Bibliography
From: j.kwisthout@observator.com (Johan Kwisthout)
Date: Wed, 10 May 2000 08:12:38 GMT
Links: << >>  << T >>  << A >>
On Tue, 9 May 2000 15:20:01 +0200, "David Brown"
<david.nospam@westcontrol.com> wrote:

>Johan Kwisthout wrote in message <3917fdf7.20732581@obsserver>...

<snip>

>>Funny, I study CS in the Netherlands and the theory of operation
>>systems (like synchronisation algorithms, multitasking theory) was
>>part of the first half of the bachalors program, dealing with
>>semaphores, mutexes, critical regions, producers/consumers, deadlocks
>>etcetera. Guess it depends on the university where you're studying...

>It certainly does.  Some universities and courses will cover things like
>Windows programming, which does not really have anything to do with CS, but
>which looks good to PHBs for later employment.  Others will teach you the
>theory of computer systems, and will cover semaphores, etc., in a manner
>which can then be applied to any system.  Part-time courses and newer
>universities tend towards the former, while older, more accademic
>universities tend towards the later (I know that's a terrible,
>unsubstantiated generalisation, but it is basically true).
>
>Dutch universities will of course tend towards the theoretical - after all,
>a high proportion of the work in synchronisation was done by Dutch
>mathematicians such as Djikstra (spelling?).

It's Dijkstra (turn the i and j the other way around). We also have
Andrew Tanenbaum hanging around here, who has done a lot of research
in OSs, being responsible for Minix, the OS that inspired Linus
Thorvalds.

>>Johan.

Article: 22493
Subject: Re: ANNOUNCE: Embedded Systems Glossary and Bibliography
From: j.kwisthout@observator.com (Johan Kwisthout)
Date: Wed, 10 May 2000 08:26:27 GMT
Links: << >>  << T >>  << A >>
On Tue, 9 May 2000 14:06:53 +0100, "PeterS" <peters AT aston desIGNS
comMERCIAL> wrote:

>Johan Kwisthout wrote in message <3917fdf7.20732581@obsserver>...

>>systems (like synchronisation algorithms, multitasking theory) was
>>part of the first half of the bachalors program, dealing with
>>semaphores, mutexes, critical regions, producers/consumers, deadlocks
>>etcetera. Guess it depends on the university where you're studying...
>>
>>Johan.
>>
>><snip>
>>
>I studied CS some 20 years ago and also covered semaphores, but don't ever
>recall the term mutex, until I came across it a couple of years ago doing
>some NT stuff. Is "mutex" purely Microsoft-speak for what everyone else
>calls a semaphore?

No. As described earlier in this thread, these are two different
things, at least in theory. A semaphore can handle multiple instances
of a resource. Say you have 10 free buffers for IO, then a semaphore
is used to assign them to processes; it holds a value from 0 to 10 to
indicate how many are used. A mutex handles one particular instance,
for example a  specific file, which one proces at a time can use
(therefore mutual exclusion). It has little to do with Microsoft.

>I don't like the use of "flag" in such specific terms - to me, a flag is
>simply an
>indication of state, with no protection mechanism for setting or resetting.

Agreed.

>BTW Didn't Dijkstra first describe semaphores, and wasn't he from the
>Netherlands?
> (not that its got anything to do with anything, much ;-)

Dijkstra (who indeed was from the Netherlands) first described an
algorithm for preventing multiple access to the same item, that he
compared with a semaphore at a railway; allowing you to pass on or
wait for the other train to pass. 

Johan.
Article: 22494
Subject: Re: ANNOUNCE: Embedded Systems Glossary and Bibliography
From: Geir Frode Raanes <geirfrs@invalid.ed.ntnu.no>
Date: 10 May 2000 08:49:15 GMT
Links: << >>  << T >>  << A >>
In comp.realtime Michael Barr <mbarr@netrino.com> wrote:

: I have put a lot of work into writing the book and creating and
: maintaining these online resources.  I hope that they will be a
: valuable contribution to the community of embedded systems hard-
: ware and software designers.

When still a student I actually bought and read your book. 
At least I think it was your book...

I had just finished a major hardware project centered around
a TMP68303 mcu with the ever popular 68000 cpu kernel.
Around it was everything I could come up with, including
DRAM, network and SCSI.  

I was then faced with the task of programming this thing. 
Of cause, beeing a student, I choosed GCC for the task.
(Still does, but that is another history) I had used GCC
on Linux before, but now I suddenly had to boot this machine
myself, handle interrupt myself, set up DMA myself and
- jiiiha - control supervisor/user space myself.  
Not to mention LibC. There is documentation on GCC, but
almost nothing on how to do this hardware fiddling.

In retroperspective, that is not a GCC issue. But green as
I was, this is where I expected to find it. Well I didn't
find it and I turned to your book for help. It helped all
right, just not where it hurt the most. Semaphores and 
message passing is fine per se, but I need to boot that
bloody system before I could start thinking about those 
issues.



So my suggestion is that you reconsider what your audience
is, or else find a new title for your book. Write a couple
of chapters on how to set up the simplest of ROM monitors.

Don't be afraid to use assembly - green programmers like me is
much more scared of C. Assembly I can see exactly what does,
while C is a black box where everything can happen. And boy,
was I scared of the dark... 

Just don't overdo the assembly part - demonstrate the principles.
Show intermixed C and assembly listings. Demystify the "crt0.s" 
issue - after all, all it does is set up a valid C runtime 
environment. All it needs to do so is some memorymapped ram 
and some rom to store itself and the main application. 
And a reset vector of cause... An issue that has linker 
written all over it. If someone would just write so, damn it!

Which leads into the issue of object code sections - 
.data, .text, .bss. These are not that scary if someone
would just point out what parts of the C code goes where and
why there will not be any leftover C constructs that does not
fit into one of these three sections. This requires a look at 
cpu addressing modes versus automatic and global C variables. 

Again - don't be afraid of explaining it in terms of hardware, 
it is exactly this understanding of the interaction between C code 
and the resulting hardware instruction issuing I was looking for.
I already knew hardware - I wanted to expand this understanding
into the abstraction of C/C++ code.

It was not until I read up on COFF that I finaly came to peace
with what the C compiler does - and more importantly where the
*linker* took over. If only someone had told me how object formats
works. Placeholder symbols is a natural consept everyone can
understand if only told about them instead of beeing patted on 
the back and told that this is beyond you. Relocation by the 
linker is then just as natural. And then it becomes just as
obvious why "crt0.s" does what it do *and nothing more.*
(Yes, it was just as big a problem for me that "crt0.s"
looked far to simple - I was looking for problems that
was just not there until I finally realised what the
the actual problem constitued of.)



To conclude - include an example hardware setup in your book - 
cpu, ram, rom and RS-232. No black boxes including on-chip ROM! 
Block diagrams plus memory maps are sufficient. I suggest the 
68K for cpu. There still is no cleaner and more academicaly 
correct cpu. And it has got a vector table - an important issue. 

Then demonstrate how to bring this machine up and running 
using GCC and Newlib. This ROM monitor should be capable 
to communicate over RS-232 to a host and possibely know 
how to download and run ready linked binaries using the 
kermit filetransfer protocol. This way you have both a 
terminal emulator and file transfer software on the host. 

There is a 68K systems book by Clemets that goes a long 
way towards this goal that I learned a lot more from than
your book. I can find the ISBN number if you want.

My two cents plus shipment and handling...

-- 
  ******************************************************
  Never ever underestimate the power of human stupidity.
  -Robert Anson Heinlein

		GeirFRS@invalid.and.so.forth
  ******************************************************
Article: 22495
Subject: appropriate ASIC Prototyping Board
From: Patrick Schulz <schulz@rumms.uni-mannheim.de>
Date: Wed, 10 May 2000 10:57:34 +0200
Links: << >>  << T >>  << A >>
Hi all,

we are targeting an PCI-based network interface ASIC prototype on VIRTEX1000(E),
but I found only very few boards which meet our requirements (PCI_64, VIRTEX1000(E)).
One of this is the DN2000k10 from the Dini Group (http://www.dinigroup.com). The problem
with this board is, that it is not compatible with the PCI LogicCore from Xilinx and I'm 
not willing to design a PCI-Interface.

Does anyone has experience with this board?
Does anyone has an idea for a appropriate board??

Thanks Patrick

-- 
Patrick Schulz (schulz@rumms.uni-mannheim.de, pschulz@ieee.org)
University of Mannheim - Dep. of Computer Architecture
68161 Mannheim - GERMANY / http://mufasa.informatik.uni-mannheim.de
Phone: +49-621-181-2720     Fax: +49-621-181-2713
Article: 22496
Subject: Spartan XCS10
From: e97bjli@thn.htu.se
Date: Wed, 10 May 2000 09:40:48 GMT
Links: << >>  << T >>  << A >>
I bought a spartan device, Xcs10 pc84.

I thought it was a 5V device. But when I use 5V Vcc all pins are high...

So, when I use 3.3V Vcc it workes okej.

But my question is...

When I look at the topp of the device, I it lookes like this....


/------------\
|xilinx
|xcs10
|PC84CKN9837
|A2014956A
|3C
\____________

Is this a 5V device or a 3.3V device?

Or is there anything wrong?

Thankful for help


Björn Lindegern



Sent via Deja.com http://www.deja.com/
Before you buy.
Article: 22497
Subject: Re: ? economical SPROM programmer for Xilinx
From: Emmanuel Lecomte <emmanuel.lecomte@free.fr>
Date: Wed, 10 May 2000 10:00:16 GMT
Links: << >>  << T >>  << A >>
Il s'agit d'un message multivolet au format MIME.
--------------F69C531C6C324827424F03BB
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit

Hello,

You could have a look to http://www.mvd-fpga.com/  -> products -> tiny prog

Greetings,

Emmanuel

Rick Filipkiewicz a écrit :

> Dan wrote:
>
> > The Data IO programmers cost $1500ish to program Xilinx SPROM. Does anyone
> > know of a more reasonably priced one ?
> >
> > Thanks Dan
>
> If you're in smallish volume why not consider the reprogrammable 1800 series
> devices from Xilinx. They are JTAG programmable. Once you've got the design
> stable you could get your SPROM distributor to program one-time ones for you.
>
> Alternatively avoid SPROMs altogether & use a standard EPROM+small CPLD to
> generate the bit stream.

--------------F69C531C6C324827424F03BB
Content-Type: text/x-vcard; charset=us-ascii;
 name="emmanuel.lecomte.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Carte pour Emmanuel Lecomte
Content-Disposition: attachment;
 filename="emmanuel.lecomte.vcf"

begin:vcard 
n:Lecomte;Emmanuel
tel;fax:05 61 06 72 60
tel;work:05 62 13 52 32
x-mozilla-html:FALSE
adr:;;;;;;
version:2.1
email;internet:emmanuel.lecomte@mvd-fpga.com
fn:Mutli Video Designs
end:vcard

--------------F69C531C6C324827424F03BB--

Article: 22498
Subject: Re: Error with Quartus for Altera APEX20K device: clock skew is greater
From: Don McCarley <mccarley@lucent.com>
Date: Wed, 10 May 2000 08:28:10 -0400
Links: << >>  << T >>  << A >>
It is not on extra effort; I don't know about assigning a clique? to a block?;
my setup and clock->out's are constrained; I will look into the Assignment
Dis-Organizer settings; I specified the clock (not global).  I did find one
thing, if I do a context help on the error, it tells me I need to add lcells to
increase data delay to the 7 problem paths, so next I will try to figure out how
to do this.  Thanks.

Don

Xanatos wrote:

> > I am a novice user of the Altera software and I am stuck with a design
> that
> > I run through Quartus 2000.03 and I get an error during compilation: clock
> > skew is greater then data delay, ciruit will not function.  If this is
>
> Is this on extra effort? Did you try to assign a clique to a block? Are your
> setup/hold times set? Check to ensure that the settings in the Assignment
> Dis-Organizer are configured properly for your design.
> Also, did you specify a global clock, or did you "specify" the clock?
>
> Check that, and if it still gives you grief, let me know.
>
> Cheers,
> Xanatos



Article: 22499
Subject: Re: HELP - what to choose?
From: husby@fnal.gov (Don Husby)
Date: Wed, 10 May 2000 13:27:54 GMT
Links: << >>  << T >>  << A >>
Rickman <spamgoeshere4@yahoo.com> wrote:
> Ray Andraka wrote:
> > ....  I haven't seen any commercial boards with Lucent, so bye
> > bye Lucent.
>
> Hey Ray! What about the PC104C31 which uses not only one, but four
> Lucent FPGA chips? I realize that our price does not fit the $300 limit
> imposed in this case, but don't say there aren't any commercial boards
> around that use the Lucent chip! :) 

A German company makes a PCI board with the Lucent FPSC (FPGA with embedded PCI)
chip:

  http://www.morethanip.com/products/or3eval/index.html

They even have a Linux driver for it.

Lucent also has an evaluation board with the PCI/FPGA chip.  At one time they
planned to sell it for $250 (it's pretty much just the chip on a board with
a few leds, switches, and daughterboard connectors).  Now they seem to have
some weird plan to loan these boards to customers.



--
Don Husby <husby@fnal.gov>             http://www-ese.fnal.gov/people/husby
Fermi National Accelerator Lab                          Phone: 630-840-3668
Batavia, IL 60510                                         Fax: 630-840-5406


Site Home   Archive Home   FAQ Home   How to search the Archive   How to Navigate the Archive   
Compare FPGA features and resources   

Threads starting:
1994JulAugSepOctNovDec1994
1995JanFebMarAprMayJunJulAugSepOctNovDec1995
1996JanFebMarAprMayJunJulAugSepOctNovDec1996
1997JanFebMarAprMayJunJulAugSepOctNovDec1997
1998JanFebMarAprMayJunJulAugSepOctNovDec1998
1999JanFebMarAprMayJunJulAugSepOctNovDec1999
2000JanFebMarAprMayJunJulAugSepOctNovDec2000
2001JanFebMarAprMayJunJulAugSepOctNovDec2001
2002JanFebMarAprMayJunJulAugSepOctNovDec2002
2003JanFebMarAprMayJunJulAugSepOctNovDec2003
2004JanFebMarAprMayJunJulAugSepOctNovDec2004
2005JanFebMarAprMayJunJulAugSepOctNovDec2005
2006JanFebMarAprMayJunJulAugSepOctNovDec2006
2007JanFebMarAprMayJunJulAugSepOctNovDec2007
2008JanFebMarAprMayJunJulAugSepOctNovDec2008
2009JanFebMarAprMayJunJulAugSepOctNovDec2009
2010JanFebMarAprMayJunJulAugSepOctNovDec2010
2011JanFebMarAprMayJunJulAugSepOctNovDec2011
2012JanFebMarAprMayJunJulAugSepOctNovDec2012
2013JanFebMarAprMayJunJulAugSepOctNovDec2013
2014JanFebMarAprMayJunJulAugSepOctNovDec2014
2015JanFebMarAprMayJunJulAugSepOctNovDec2015
2016JanFebMarAprMayJunJulAugSepOctNovDec2016
2017JanFebMarAprMayJunJulAugSepOctNovDec2017
2018JanFebMarAprMayJunJulAugSepOctNovDec2018
2019JanFebMarAprMayJunJulAugSepOctNovDec2019
2020JanFebMarAprMay2020

Authors:A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Custom Search